Molecular Phylogenetics and Evolution 77 (2014) 116–125

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev

Phylogenetic relationships within the snapping shrimp genus Synalpheus (Decapoda: Alpheidae) Kristin M. Hultgren a,b,⇑, Carla Hurt c,e,1, Arthur Anker d a

Biology Department, Seattle University, Seattle, WA 98122, USA National Museum of Natural History, Smithsonian Institution, Washington, DC 20013, USA c Department of Biology, University of Miami, Coral Gables, FL 33143, USA d Tropical Marine Science Institute, National University of Singapore, Singapore 119227, Republic of Singapore e Department of Biology, Tennessee Tech University, Cookeville, Tennessee, 38505, USA b

a r t i c l e

i n f o

Article history: Received 20 July 2013 Revised 9 March 2014 Accepted 11 March 2014 Available online 28 March 2014 Keywords: Synalpheus Phylogeny Alpheidae Snapping shrimp Cryptic species

a b s t r a c t The snapping shrimp genus Synalpheus (Alpheidae) is one of the most speciose decapod genera, with over 160 described species worldwide. Most species live in symbiotic relationships with other marine organisms, such as sponges, corals and crinoids, and some sponge-dwelling species have a highly organized, social structure. The present study is the first worldwide molecular phylogenetic analysis of Synalpheus, based on >2200 bp of sequence data from two mitochondrial (COI and 16S) and two nuclear (PEPCK and 18S) loci. Our molecular data show strong support for monophyly of three out of six traditionally recognized morphology-based species groups: the S. brevicarpus, S. comatularum and S. gambarelloides groups. The remaining three species groups (S. paulsoni, S. neomeris and S. coutierei groups) are non-monophyletic in their current composition and will need to be either abandoned or taxonomically redefined. We also identified potential cryptic species of Synalpheus in our dataset, using intraspecific and interspecific sequence variation in COI from the taxonomically well-studied S. gambarelloides group to establish a genetic divergence threshold. We then used both genetic divergence and tree-based criteria (reciprocal monophyly) to identify potential cryptic species in the remaining taxa of the genus. Our results suggest the presence of multiple cryptic lineages in Synalpheus, underlining the need for more integrative taxonomic studies—including morphological, ecological, molecular, and color pattern data—in this biologically interesting genus. Ó 2014 Elsevier Inc. All rights reserved.

1. Introduction Species-rich marine groups present a particular challenge for both taxonomy and systematics, as they often contain cryptic taxa that are difficult to identify using traditional, morphology-based criteria (Knowlton, 1986, 1993, 2000; Hebert et al., 2003; Witt et al., 2006; Barber and Boyce, 2006; Mathews, 2006). Diagnosing and defining species in these groups, and examining their phylogenetic relationships, is critical for our understanding of evolution and diversity of tropical marine invertebrate communities. Shrimps of the family Alpheidae represent one of the most diverse groups of marine decapod crustaceans, with 45 genera and over 600 described species worldwide (Chace, 1972, 1988; Anker ⇑ Corresponding author at: Biology Department, Seattle University, Seattle, WA 98122, USA. E-mail address: [email protected] (K.M. Hultgren). 1 Current address: Department of Biology, Tennessee Tech University, Cookeville Tennessee, 38505, USA. http://dx.doi.org/10.1016/j.ympev.2014.03.008 1055-7903/Ó 2014 Elsevier Inc. All rights reserved.

et al., 2006; De Grave and Fransen, 2011). Alpheid shrimps are among the most abundant decapods in tropical and subtropical, shallow-water habitats, in particular on coral reefs (Pearse, 1932; Ruetzler, 1976; Felder and Chaney, 1979). Many alpheids live in permanent association with a variety of other marine invertebrates and gobiid fishes (Banner and Banner, 1975; Bruce, 1976; Anker et al., 2005; Rios and Duffy, 2007; Karplus and Thompson, 2011). The genus Synalpheus Spence Bate, 1888 is the second-largest genus in the family, with over 160 described species worldwide (Banner and Banner, 1975; Chace, 1988; De Grave and Fransen, 2011). The ecological diversity of Synalpheus, which includes multiple instances of symbioses with sponges, echinoderms and cnidarians (Fig. 1), and several modes of social organization, has made this genus an attractive model system for studies of speciation, host specialization and evolution of sociality (Duffy, 1992, 1996; VandenSpiegel et al., 1998; Duffy and Macdonald, 2010; Hultgren and Duffy, 2012). However, most of the research in the genus Synalpheus has focused on a single clade of 45 described Caribbean sponge-dwelling species in the Synalpheus gambarelloides species group, and examining the

K.M. Hultgren et al. / Molecular Phylogenetics and Evolution 77 (2014) 116–125

117

Fig. 1. Diversity of host ecology and coloration in the snapping shrimp genus Synalpheus: A – S. neomeris, an Indo-West Pacific species associated with soft corals (Dendronephthya sp.); B – S. stimpsonii, an Indo-West Pacific species complex associated with comasterid feather stars; C – S. antillensis, a West Atlantic species free-living in coral rubble; D – S. modestus, an Indo-West Pacific species, free-living or commensal of sessile invertebrates; E – S. charon, an Indo-Pacific species associated with pocilloporid corals; F – S. dardeaui, a West Atlantic sponge-dwelling species (here on partly cut-open Lissodendoryx colombiensis). Photographic credits: A, B, D – Ned Deloach; C, E, F – Arthur Anker.

evolution of specialized host use across the entire genus requires a robust phylogeny that spans the worldwide diversity of this group. Due in part to the spectacular diversity of Synalpheus in tropical marine habitats, the taxonomy of many described species remains in an unsatisfying state. This is especially true for numerous IndoWest Pacific and East Pacific taxa, some with problematic synonyms, subspecies and varieties (De Man, 1888; Coutière, 1905, 1908, 1909, 1921; De Man, 1911). Furthermore, almost nothing is known about the phylogenetic structure within this genus worldwide (Banner and Banner, 1975; Anker and De Grave, 2008; Hermoso-Salazar et al., 2008). Coutière (1908, 1909) subdivided the genus into six informal species groups: S. brevicarpus, S. biunguiculatus (later changed to S. coutierei), S. comatularum, S. laevimanus (later changed to S. gambarelloides), S. neomeris, and S. paulsoni groups. Subsequent taxonomic treatments (Banner and Banner, 1975) concluded that only three of these species groups (S. brevicarpus, S. comatularum, and S. gambarelloides groups) had enough morphological support to be taxonomically useful (see also Hermoso-Salazar et al., 2008). The S. brevicarpus group contains approximately a dozen species (both sponge-symbionts and non-symbiotic, Fig. 1C), some currently under description (Anker, unpublished data), distributed exclusively in the eastern Pacific and western Atlantic. The S. comatularum group includes at least 10 described species, all found in the tropical parts of the Indo-West Pacific; most (if not

all) of its members are associates of crinoids (Fig. 1B). The very large S. gambarelloides species group (>70 described species) is distributed worldwide, although the vast majority of species occur in the tropical western Atlantic. It is by far the best-studied group of Synalpheus, in terms of ecology and phylogenetics. All members of the S. gambarelloides group are ecologically quite homogeneous, dwelling exclusively in the interior canals of sponges (Banner and Banner, 1975; Dardeau, 1984; Chace, 1988; Ríos and Duffy, 2007). Molecular studies have consistently indicated strong support for the monophyly of this group, although prior phylogenies sampled only a few taxa outside of the S. gambarelloides group (Duffy et al., 2000; Morrison et al., 2004; Hultgren and Duffy, 2011). Based on these data, Ríos and Duffy (2007) erected a new genus, Zuzalpheus, for the S. gambarelloides group. However, Anker and De Grave (2008) pointed out that Zuzalpheus was separated from Synalpheus based on minor and ambiguous morphological differences, and this rendered the rest of the genus Synalpheus paraphyletic, based on the phylogeny proposed by Morrison et al. (2004). The resolution of relationships within Synalpheus, and possible establishment of morphologically defined subgenera (including Zuzalpheus), requires a much more extensive and worldwide sampling of taxa from all six of Coutière’s species groups worldwide. Identifying and describing cryptic species complexes is a major challenge for assessing worldwide species diversity in Synalpheus. For example, since Dardeau’s (1984) work on the western Atlantic

118

K.M. Hultgren et al. / Molecular Phylogenetics and Evolution 77 (2014) 116–125

species of the S. gambarelloides group, many cryptic species complexes have been identified and split into several species; as a result, the total number of described species in the S. gambarelloides group has more than doubled over the last two decades, from 19 to currently 44 (summarized in Hultgren and Duffy, 2011). In several cases, it may be difficult to use morphological characters alone to accurately delimit species; in this group, molecular sequencing has assisted in discriminating between morphologically similar species (Hultgren et al., 2010), and has strongly supported the morphology-based species concepts (Morrison et al., 2004; Hultgren and Duffy, 2011). Overall, a more integrative approach is needed, which includes molecular data (e.g., COI barcoding gene), as well as color patterns, ecology (e.g., hosts for symbiotic species), and traditional morphological characters. In many animal groups, molecular data are increasingly being used to identify potential cryptic species (Hebert et al., 2004; Witt et al., 2006; Barber and Boyce, 2006; Oliver et al., 2009), with the 650 bp 50 region of the mitochondrial cytochrome oxidase I (COI) gene the generally accepted DNA barcode marker (Hebert et al., 2003; Goldstein and Desalle, 2010). Some workers have utilized a delimitation threshold of ten times greater than intraspecific distance (Hebert et al., 2004; Witt et al., 2006), and more recently the ‘‘gap’’ between intraspecific and congeneric heterospecific genetic distance in COI has been utilized as a threshold for species delimitation (Lefebure et al., 2006; Costa et al., 2007; Radulovici et al., 2009; Del-Prado et al., 2010; Goldstein and Desalle, 2010; Puillandre et al., 2012). Other workers advocate a tree-based approach, specifically using reciprocal monophyly as a criterion for species delimitation (reviewed in Goldstein and Desalle, 2010), and many studies use some combination of tree-based and genetic barcoding approaches to identify potential cryptic species (Barber and Boyce, 2006; Xavier et al., 2010; Murray et al., 2012). Ultimately, these potential cryptic species can be confirmed by future studies through traditional taxonomic and coalescent-based species tree inference methods. The main goal of the present study is to propose the first worldwide phylogeny of the genus Synalpheus, including representatives from all six informal species groups (Coutière, 1909; Banner and Banner, 1975) and from all four oceanic provinces (East Atlantic, West Atlantic, East Pacific, and Indo-West Pacific). We constructed phylogenetic trees using molecular sequence data from four loci: two nuclear loci and two mitochondrial loci, including the 50 region of the mitochondrial COI gene used for genetic barcoding, to assess monophyly of each of the six species groups, and the relationships among these groups. Within the taxonomically well-studied S. gambarelloides group, we calculated intraspecific and interspecific divergence in COI, used these data to establish a genetic distance threshold, and used a combination of tree-based and genetic distance criteria to identify potential cryptic species in the remaining taxa of Synalpheus. This study provides the foundation for further taxonomic work exploring diversity and clarifying phylogenetic relationships within the genus Synalpheus.

2. Material and methods 2.1. Taxon sampling and species identification Sequence data were generated from specimens of Synalpheus and Alpheus (latter used as outgroup) collected primarily over the last decade (2001–2010), including numerous specimens collected and processed by two of the authors (AA and KH, see Supplementary data 1). We also used 16S sequence data from prior studies (Morrison et al., 2004; Hultgren and Duffy, 2011) and sequence data (COI, 16S, and 18S) generated by the Smithsonian Institution for the Barcode of Life project (archived by the Barcode of Life Data

Systems, or BOLD). Specimens were collected intertidally and subtidally (scuba diving or snorkeling), usually by breaking up coral rubble or examining interior of sponges; some were found under rocks or on crinoids and living corals. Most specimens were photographed alive in the field and preserved in 80–95% ethanol. Species identifications were made by AA and KH based on existing species descriptions (e.g., Banner and Banner, 1975; Ríos and Duffy, 2007; Anker and Tóth, 2008). Collection locations, host/habitat information, voucher locations, and additional taxonomic information are summarized in Supplementary data 1; information for selected species was also deposited in the Barcode of Life Data System website (BOLD, www.barcodinglife.org). 2.2. DNA/RNA extraction, sequencing, and alignment For most specimens, RNA was extracted using the SV Total RNA Isolation System (Promega, Madison WI) using a modified protocol that preserved some genomic DNA (Regier, 2008). For a subset of specimens, we used genomic DNA that was extracted using the Qiagen tissue protocol on a Biosprint 96 workstation. Sequencing was done at the University of Miami core sequencing facility or through the Smithsonian’s DNA barcoding project (Laboratories of Analytic Biology, hereafter LAB). We sequenced individuals for four loci: the 50 end of the mitochondrial cytochrome oxidase I gene used for genetic barcoding (COI); partial sequences of the nuclear large ribosomal subunit 18S (18S); partial sequences of the mitochondrial rRNA molecule 16S (16S); and partial sequences of the nuclear phosphoenolpyruvate carboxykinase gene (PEPCK). 18S and 16S were amplified directly; for COI, we used both direct amplification using degenerate primers and in some cases a RTPCR protocol, modified from Hurt et al. (2009), to reduce the risk of amplifying of nuclear pseudogenes (Williams and Knowlton, 2001); for interior and exterior primers, see Table 1. The RT-PCR protocol was also used to amplify the nuclear PEPCK locus in order to reduce the risk of amplifying non-functional copies of the gene and to allow us to predict the size of the desired amplicon. Briefly, we synthesized cDNA from RNA extractions using MuLV reverse transcriptase (Applied Biosystems), RNase inhibitor (Applied Biosystems), and a sequence-specific reverse primer. This cDNA was used as a template in a PCR reaction that utilized a forward primer and PCR/thermocycler conditions suggested by Regier (2008). PCR products of the correct size were excised from a 1% low-melt agarose gel, and extracted using the GELase protocol (Epicentre Biotechnologies), and sequenced on an ABI Prism 3130xl sequencer. Forward and reverse sequences from all loci were visually checked and trimmed using the program SEQUENCHER v4.8 (Gene Codes corporation). Each individual locus was aligned using the default parameters on the program Muscle v3.8 (Edgar, 2004), and we used MacClade 4.08 (Maddison and Maddison, 2005) to translate coding loci (PEPCK and COI) into protein sequence to check for stop codons, which may indicate the presence of pseudogenes previously reported in Alpheus (Williams and Knowlton, 2001; Buhay, 2009). No stop codons were detected in any of the coding sequences. We also examined individual gene trees to check for unusual species placements that could indicate pseudogenes or multiple coding copies of the loci used in the study. In the case of PEPCK, we occasionally amplified multiple copies of the PEPCK locus (sometimes from the same individual) in individuals of Synalpheus. However, different loci formed well-supported clades that were distinct outgroups to the most commonly amplified locus, and were easily distinguished and removed. In most cases, sequences from the second locus did not contain stop codons, suggesting they may not necessarily be pseudogenes but possibly multiple functional copies. We trimmed sequences to exclude the trailing ends for COI (658 bp region utilized), 16S (603 bp), and PEPCK (519 bp). Although alignment

K.M. Hultgren et al. / Molecular Phylogenetics and Evolution 77 (2014) 116–125

119

Table 1 Primers used in analysis, primer source, sequence generation, and amplification strategy; Smithsonian LAB indicates sequences generated at the Laboratory of Analytical Biology through the DNA barcoding project. Locus/primer name

Sequence 50 –30

COI

jgLCO1490 jgHCO2198 jgHCO_2198 jgLCO_1490

50 -GMA TAG TAG GMA CRG CYC TNA-30 50 -YCC TGT GAA TAG GGG GAA TC-30 50 TAI ACY TCI GGR TGI CCR AAR AAY CA 30 50 TIT CIA CIA AYC AYA ARG AYA TTG G 30

PEPCK

PEPCK-3F PEPCK-454R

50 -GCT ATG AAA ACC GTC CTT TCC-30 50 -TGC TGT AGG TAG TGG CCA AA-30

Alpheus/Synalpheus Alignment Alpheus/Synalpheus Alignment

RT-PCR (U. Miami)

18S

18S-Y 18S-A SR7_18Sr EukF_18Sf

50 -CAG ACA AAT CGC TCC ACC AAC-30 50 -AAC CTG GTT GAT CCT GCC AGT-30 50 -GTTCAACTACGAGCTTTTTAA-30 50 -AACCTGGTTGATCCTGCCAGT-30

Apakupakul et al. (1999) Medlin et al. (1988) Vilgalys and Bun (1994) PNAS Sands et a. BMC Ecol. 2008

Direct PCR (U. Miami)

16S-1472R 16S-ar 16Sbr

50 -AGA TAG AAA CCA ACC TGG-30 50 -CGC CTG TTT ATC AAA AAC AT-30 50 -CCG GTC TGA ACT CAG ATC ACG T-30

Crandall and Fitzpatrick 1996 Palumbi et al. (1991)

Direct PCR (16Sar/16S1472R or 16Sar/16Sbr) (U. Miami) Direct PCR (16Sar/16Sbr)(Smithsonian LAB)

16S

Primer source

Amplification/location Direct PCR or RT-PCR (U. Miami) Direct PCR (Smithsonian LAB)

of the 16S locus was fairly straightforward, some portions of the 18S locus were highly divergent among different described species and difficult to align. We used the program GBlocks v0.9 (Castresana, 2000; Talavera and Castresana, 2007) to exclude ambiguous areas of the alignment using relaxed gap selection criteria (allowed gap positions = all) suggested from simulation studies (Talavera and Castresana, 2007), yielding a 457-bp region for analysis near the 50 end of our original sequences. As there are few well-tested methods for incorporating gap information into model-based phylogenetic methods (Phillips et al., 2000; Simmons et al., 2007), we treated all gaps as missing data. We used the program MrModeltest v.2.3 (Nylander, 2004) to assess different models of nucleotide substitution for each of the three loci, and selected the model of molecular evolution with the best fit to the data using the Akaike Information Criterion.

2.3. Tree construction We constructed both single-locus gene trees and combined data trees from the sequence data. We first constructed individual gene trees to examine congruence among the four genetic loci we used (COI, 16S, PEPCK, and 18S). Next, we constructed a combined-data tree using only individuals that were sequenced for COI, PEPCK, and 18S (the three loci sampled most frequently); this ‘‘complete data’’ tree maximized sequence coverage. However, as extraction of nuclear loci using the RT-PCR method was difficult from older samples (i.e., preserved in ethanol for >5 years), portions of molecular datasets were missing for some taxa; for example, none of the S. comatularum complex (and only one species from the S. coutierei complex) were included in the complete-data tree. As simulation studies suggest that utilization of taxa with at least P50% of the data matrix present can improve the accuracy of trees (Wiens, 2005, 2006), we also constructed a tree that maximized taxon sampling by including all specimens sequenced for COI mtDNA and at least one other locus. This ‘‘complete-taxa’’ tree included specimens sequenced for 2-4 loci, excluding individuals with 2 individuals sampled. We used MEGA v.4 (Tamura et al., 2007) to calculate nucleotide divergences using the Kimura two-parameter (K2P) distance model (Kimura, 1980). Interspecific distances were calculated to compare to published crustacean congeneric distance values for COI (Lefebure et al., 2006; Costa et al., 2007), and sister species comparisons were used to calculate minimum interspecific genetic distance (Meier et al, 2008). Mean sister species divergence was 8.1% (n = 7 pairs), while mean intraspecific divergence was 1.02% (n = 16; see Results). We used two criteria to identify

120

K.M. Hultgren et al. / Molecular Phylogenetics and Evolution 77 (2014) 116–125

potential cryptic species in the remaining species groups of Synalpheus. First, genetic divergence in COI between a potential cryptic species and its closest sister species taxa or group had to exceed 10.2%, i.e. ten times mean intraspecific divergence. This percentage also exceeded minimum interspecific distance—i.e., mean sister-species divergence (8.1%)—in the S. gambarelloides group. Second, multiple individuals of a potential cryptic species had to be reciprocally monophyletic in both combined-data (completetaxa tree) and COI gene trees. In this study, the four undescribed species in the S. gambarelloides group are distinguishable by several morphological characters and are currently under taxonomic study. 3. Results 3.1. Alignment and sequence data We obtained mitochondrial COI sequence data from 196 individuals representing 93 different species (here including both described and potential cryptic species), which is about 60% of the current described species diversity of Synalpheus. For PEPCK, we obtained sequence data from 91 individuals (67 species); for 16S, we obtained sequence data from 176 individuals (87 species); for 18S we obtained sequence data from 105 individuals (60 species). The complete-data tree (Fig. 2) consisted of 63 total individuals representing 48 species, while the complete-taxa tree (Fig. 3) consisted of 133 individuals representing 88 species (30 species with 2 loci, 54 species with 3 loci, 26 species with 4 loci). Best-fit nucleotide substitution models calculated from MrModeltest were as follows: COI, 18S, and 16S: GTR + I + G; PEPCK: HKY + I + G. 3.2. Phylogenetic trees Although we constructed separate gene trees for each molecular locus to examine congruence among all four datasets, only combined-data trees provided resolution of phylogenetic relationships among clades with high Bayesian posterior probabilities or bootstrap values. Although ML and Bayesian trees resolved the same major clades (Table 2), Bayesian trees provided more information about relationships among these clades. Therefore, we present the Bayesian topologies with Bayesian posterior probabilities (bpp), along with ML bootstrap (bs) values on clades resolved with both methods (all support values are converted to percentages). Bayesian gene trees for COI 16S, 18S, and PEPCK are presented in in Supplementary Figures 2–5, respectively. In general, four major clades of Synalpheus were identified in both combined-data trees, but the complete-data tree tended to provide higher support for these clades (Table 2, Figs. 2 and 3). First, monophyly of the S. brevicarpus group was strongly supported in both combined-data trees using all methods (bpp, bs = 100), and the S. gambarelloides group was supported in combined Bayesian analyses (bpp = 100), but was not well supported in ML analyses (bs = 28–60). Second, there was strong support in both combined-data trees (bpp, bs P 97) for ‘‘Clade I,’’ which consisted of eight species (described and potential cryptic species) from the S. paulsoni group, primarily from the eastern Pacific (six species), with one species each from the western Atlantic and the eastern Atlantic. This clade was also supported in all individual gene trees (bpp P 97) except 18S, and was recovered as a sister clade to the S. comatularum clade in the complete-taxa tree. Third, there was also strong support in both combined-data trees (bpp = 100, bs = 80–98) for ‘‘Clade II,’’ which consisted of species (sampled worldwide) currently assigned to the S. coutierei, S. paulsoni, and S. neomeris groups (Coutière, 1909; De Man, 1911; Banner and Banner, 1975). Finally, the S. comatularum clade had strong

support in the complete-taxa tree (bpp = 99); this group was not included in the complete-data tree due to insufficient genetic sampling. Combined-data trees and the aligned data matrix are available on Treebase: (http://purl.org/phylo/treebase/phylows/ study/TB2:S14730). 3.3. Cryptic species delimitation In the S. gambarelloides group, mean intraspecific divergence (mean ± standard error) was 1.02 ± 0.92% (n = 16); mean sisterspecies divergence was 8.14 ± 1.41% (n = 7); mean interspecific divergence was 19.57 ± 0.21% (n = 496 comparisons). Potential cryptic species were identified using a COI sequence divergence threshold of 10.2% (10 intraspecific divergence), and all potential cryptic species (with >1 individual sampled) were reciprocally monophyletic in both complete-taxa (Fig. 3) and COI (Supplementary Fig. 2) trees. Using these criteria, we detected potential cryptic species in the S. comatularum clade (2 species), Clade I (3 species), the S. brevicarpus clade (5 species), and Clade II (10 species). The species coverage across the six species groups, expressed as % of currently known diversity (Anker, 2001a; De Grave and Fransen, 2011), was 17% for the S. coutierei group, 25% (S. comatularum), 65% (S. gambarelloides), 68% (S. neomeris), 57% (S. paulsoni), and 100% (S. brevicarpus).

4. Discussion The first worldwide phylogenetic study of the diverse snapping shrimp genus Synalpheus demonstrates strong support for the monophyly of three out of six species groups originally established by Coutière (1909): the S. gambarelloides, S. brevicarpus, and S. comatularum groups. However, the phylogenetic relationships among these groups, and the other two major clades recovered in the combined-data trees, were generally less resolved. 4.1. Monophyly of Coutière’s species groups The S. gambarelloides, S. brevicarpus, and S. comatularum groups were the only three species groups consistently recovered as monophyletic in combined-data trees (Figs. 2 and 3, Table 2) and in COI and 16S gene trees (Supplementary Figs. 2–3). These data suggest a strong congruence between our molecular topologies and morphology-based taxonomic hypotheses, as these are the only three species groups deemed taxonomically coherent by Banner and Banner (1975). Importantly, two of these groups are also characterized by obligate symbiotic associations: all of the members of the S. gambarelloides group dwell exclusively in sponges (Ríos and Duffy, 2007; Fig. 1F), whereas members of the S. comatularum group are thought to be primarily (if not exclusively) associated with crinoids (Banner and Banner, 1975; Fig. 1B). As predicted by Banner and Banner (1975), there was no support in our study for monophyly of the S. neomeris, S. paulsoni and S. coutierei groups, as defined by Coutière (1909) and De Man (1911). In our study, members from these groups are distributed among Clades I and II (Figs. 2 and 3). The smaller Clade I is formed by several eastern Pacific and Atlantic species that are more or less closely related to S. apioceros, a species originally assigned to the S. paulsoni group by Coutière (1909). The much larger Clade II is a morphologically, ecologically and biogeographically heterogeneous assemblage of members of the S. neomeris and S. coutierei groups, as well as the remaining members of the S. paulsoni group sensu Coutière (1909). However, Clade II contains several biogeographically constrained sub-clades with species from the Indo-West Pacific (S. neptunus / S. coutierei complex, S. neomeris

K.M. Hultgren et al. / Molecular Phylogenetics and Evolution 77 (2014) 116–125

121

Fig. 2. Bayesian consensus tree of all individuals sequenced for 18S, PEPCK, and COI (complete-data tree); specimen number and collection locality are included. Synalpheus species groups are indicated by different colors; major clades indicated by vertical bars. Numbers by each node indicate Bayesian posterior probability values (bpp)/bootstrap support from maximum likelihood analyses (bs), expressed as percentages. A ‘‘–‘‘ indicates that the clade is present but bs or pp values < 50; ‘‘n’’ indicates clade was not present in ML trees. Values below the level of species are omitted. A ‘‘’’ indicates potential cryptic species identified using COI sequence divergence data. Scale bar indicates number of substitutions/site in the Bayesian analysis. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

complex, Fig. 1A), Indo-Pacific (S. charon complex, Fig. 1E), or West Atlantic – East Pacific (S. fritzmuelleri complex). Prior to our study, the only formal phylogenetic analysis of the genus Synalpheus was a study focused on the American members of the S. paulsoni group, using a morphological dataset of 51 characters (Hermoso-Salazar et al., 2008). The authors of this study concluded that only two species groups should be recognized in Synalpheus: the S. paulsoni group and the S. gambarelloides group. However, their taxa set was heavily biased towards American species, especially those assigned to the S. paulsoni group. The large S. gambarelloides group was represented only by two East Pacific taxa, S. occidentalis and S. mulegensis. The S. paulsoni group sensu lato, as recovered as monophyletic by Hermoso-Salazar et al. (2008), included an array of morphologically different species, such as S. fritzmuelleri, S. apioceros, S. charon and S. digueti (the latter is a member of the S. brevicarpus group). In contrast, our molecular analysis provided no support for a monophyletic S. paulsoni group.

For instance, S. charon and S. fritzmuelleri were formed two distinct lineages within Clade II, and S. apioceros clustered with several other species in Clade I. Both our molecular study and the morphological study by Hermoso-Salazar et al. (2008) recovered a clade that included S. fritzmuelleri, S. nobilii, S. sanlucasi, and S. hemphilli (within Clade II in our trees). However, Hermoso-Salazar et al. (2008) also placed S. bannerorum in this clade, while previous molecular work (Morrison et al., 2004) and our study recovered S. bannerorum and S. dominicensis as transithsmian sister taxa most closely related to the S. gambarelloides group. The remaining topology of the S. paulsoni group in Hermoso-Salazar et al. (2008) was mostly not supported by our molecular analysis, with S. townsendi and S. spinifrons recovered in Clade II, and S. lani, S. stylopleuron, S. sanjosei, S. apioceros, and S. peruvianus recovered in Clade I. The phylogenetic analysis of Hermoso-Salazar et al. (2008) underlines the enormous difficulties of purely morphological species-level analyses in rapidly evolving groups.

122

K.M. Hultgren et al. / Molecular Phylogenetics and Evolution 77 (2014) 116–125

Fig. 3. Bayesian consensus tree (complete-taxa tree), constructed with 18S, PEPCK, 16S, and COI sequence data. Major species groups, major clades, scale bar, and other abbreviations as in Fig. 2. Numbers by each node indicate either Bayesian posterior probability values (bpp) or bootstrap support (bs) from ML analyses, expressed as percentages; bpp/bs (bpp and bs P 98 are indicated by a dot).

123

K.M. Hultgren et al. / Molecular Phylogenetics and Evolution 77 (2014) 116–125

Table 2 Phylogenetic support for different species groups in Synalpheus, for Bayesian (BY) and maximum likelihood (ML) trees constructed from different character sets (gene trees) and taxon sets (combined trees).

brevicarpus comatularum coutierei gambarelloides neomeris paulsoni Clade I Clade II

COI (Supp. Fig. 2)

16S (Supp. Fig. 3)

18S (Supp. Fig. 4)

PEPCK (Supp. Fig. 5)

Complete-data tree (Fig. 2)

BY

BY

BY

BY

BY

ML

Complete-taxa tree (Fig. 3) BY

ML

98 97 – – – – 97 –

100 100 – – – – 100 –

– – – – – – – –

99 n/a – – – – 100 –

100 n/a – 100 – – 100 100

100 n/a – 60 – – 99 98

100 100 – 100 – – 100 100

100 100 –

Phylogenetic relationships within the snapping shrimp genus Synalpheus (Decapoda: Alpheidae).

The snapping shrimp genus Synalpheus (Alpheidae) is one of the most speciose decapod genera, with over 160 described species worldwide. Most species l...
2MB Sizes 1 Downloads 3 Views